The Unspecified Drive

Deep space propulsion is, unsurprisingly, a major concern of this blog. I regularly specify the performance of interplanetary craft fitted with some form of high specific impulse drive. Sometimes I describe it as a nuclear electric or solar electric drive, sometimes simply as electric, often not even that much. Sometimes, especially when discussion takes us to the wide open spaces beyond Jupiter, I allude to fusion.

Since I got myself in a bit of hot water, or some more exotic (and much hotter) coolant, by some snide remarks about fission power plants, a few comments on deep space propulsion are in order.

First of all it is not the main barrier to widespread interplanetary travel. That would be the sheer amount of costly design engineering needed to build a fleet of prototype spacecraft, followed by the cost of getting them all into space.

But once we are up there, how we get around is an important concern. The current limit to human space missions is about six months, beyond which the health consequences of prolonged microgravity become severe. Longer missions require a spin hab, adding cost and complexity. Even with spin habs, radiation and ordinary human factors limit practical mission duration to a couple of years or so.

Within these constraints we could reach Mars with chemfuel (and a spin hab), but the Hohmann round trip to the main asteroid belt is two and a half years, without any stay time at the destination, while to Jupiter and back is five and a half years. This is too long for regular human travel.

So for a human interplanetary presence we need fast orbits. These are above my math pay grade to calculate, but a klugewerks of flat space modeling, sketching orbits, interpolation, and sheer guesswork indicates that reaching Mars in three months or Jupiter in a year calls for a mission delta v in the range of about 30-100 km/s, and therefore some form of high specific impulse drive. Even NERVA style nuclear thermal rockets - the classic Atomic Rockets that gave the website its name - fall short of this requirement.

The time honored high specific impulse drive in science fiction is ion propulsion, used in real life to send the Dawn mission to Ceres and Vesta, but not suited to much larger human-carrying spacecraft. To a great many people, however, 'ion drive' is more or less synonymous with electric drive in general.

The most likely such drive for human missions appears to be some form of plasma jet. Unlike ion drive this is a thermal drive: The plasma has a meaningful temperature - and it is extremely hot. But the thrust chamber is a magnetic field, so it won't melt. Only the gizmos that produce the field are exposed, and they don't get up close and personal with the plasma. They and their supporting struts must have heat shielding, forming a 'lantern' structure.

(The strictly technical term for this drive is electrothermal magneto-plasma propulsion - doesn't that sound exactly like classic Trek technobabble? "I've engaged the electrothermal magneto-plasma thrusters, Keptain - she canna take much more!")

So far plasma drive has gone no further than the laboratory bench, but there don't seem (yet) to be any serious problems in scaling it up to be suitable to large spacecraft. Like many forms of electric drive it has no inherent exhaust velocity and therefore no fixed specific impulse. At least in principle these drives can be configured either to expel a relatively large flow of relatively (very relatively!) cool plasma at lower velocity, or a smaller quantity of hotter plasma at higher velocity.

The effect is very closely analogous to gearing; these drives can be set for a higher acceleration and lower specific impulse or vice versa. VASIMR is supposed to achieve this not only in principle but in engineering practice, permitting clever tweaking of engine settings to get the optimum performance in each phase of flight.

For all of its advantages, electric drive has one essential drawback. It does not produce its own energy, as chemfuels do, or even use a reactor directly to heat the propellant, as nuke thermal does. It must be plugged into an external electric power supply. This is seriously inconvenient, because it takes a lot of electric power, tens to hundreds of megawatts, to drive a big, human carrying ship even at milligee acceleration.

For travel in the inner system I am partial to solar electric power. It hums along quietly with little fuss and practically no moving parts. But the butterfly's wings must be enormous, a hectare for every few megawatts, and extremely light. Even milligee forces may be problematic when the wing structure is that big and that light. And solar electric fades rapidly with distance from the Sun, unsuitable for travel beyond Mars.

For the asteroid belt and Jupiter the practical alternative is nuclear electric drive, which was the cause of my original grump. All vivid if misleading imagery of clanking steam engines aside, nuclear power plants are heavy, filled with complex plumbing that must operate for months under fiercely hostile conditions, and produce two or three times their useful output in waste heat, which must be got rid of through large radiators with their own demanding plumbing.

That eerie green glow is produced by the disintegration of money.

There is an upside to all this downside: Ships with nuclear or solar electric drive have plenty of juice at the main switchboard, making these drives, especially nuke electric, well suited to laser stars. All you need is the laser installation; the power supply is already provided, and you can zap away as long as you want to hold down the trigger.

But the general messy inconvenience of carrying around a naval-equivalent fission power reactor accounts for much of the appeal of fusion. In principle, and popular imagination, fusion is an ideal power source for a plasma drive, because the fusion plasma and the thrust plasma can be one and the same. VASIMR in fact is a byproduct of fusion research; in a conceptual sense it is a derated fusion drive.

Fusion in practice could turn out to be another matter. What else is new? The easiest fusion reactions to sustain (and we can't yet fully sustain any of them) release most of their energy as neutrons, useless for propulsion, but - irony alert - suitable for heating a steam boiler.

On the other hand, fusion propulsion is in some respects simpler than fusion power for earthly energy needs. It does not need to be an economical means of producing electric power. In fact to serve as a drive it need not produce any electric power at all, though any fusion drive would likely produce some 'bleed' power.

There are alternatives to fusion, all about as speculative as fusion itself. Orion is arguably the least speculative of the bunch, though the organ music and black cape factor has pretty much overshadowed the actual technical challenges of building a spacecraft that must nuke itself thousands of times, at close range, in the course of normal operation. (Those have to be some badass shock absorbers!)

But on the whole the specific technical details of a high specific impulse drive matter surprisingly little. What matters is how heavy the thing is, relative to the thrust power it puts out. The benchmark here is is a specific power output of roughly 1 kW/kg, or a megawatt per ton, for the full drive installation including thrusters, power supply, and waste heat radiators. (For a complete drive bus add propellant tankage and keel structure; mate it all to a payload to get a ship.)

Example: Suppose a 100 MW, 100 ton VASIMR style drive engine. With exhaust velocity tuned to 75 km/s, specific impulse near 7500 seconds, propellant mass flow is 36 grams/second, producing about 2.7 kN of thrust, enough to push a 500 ton ship at just over a half a milligee, gradually increasing as propellant is burned off. If half the departure mass is propellant (250 tons, plus 100 tons for the drive, leaving 150 tons for tankage, structures, and payload), mission delta v is just over 50 km/s. Full power burn duration is about 80 days. This broadly corresponds to the requirement for a fast, three month orbit to Mars.

Tune the same drive to an exhaust velocity of 150 km/s, specific impulse near 15,000 seconds. Propellant mass flow falls to about 9 grams/second, producing 1.3 kN of thrust, pushing the ship at a quarter of a milligee. With the same mass proportions our ship has a mission delta v of just over 100 km/s and full power burn duration of 11 months, approximating a one year trip to Jupiter.

Improving on this performance will not be easy. To reduce travel time on semi-brachistochrone orbits with prolonged burns you must reach a higher peak speed in less time, and must therefore increase both thrust and specific impulse. In the flat space approximation, drive power increases as the inverse cube of travel time - that is, you need eight times the drive power output to cut travel time in half.

The good news, such as it is, is that this also works the other way. An early generation drive with a more modest 250 W/kg power output can still take a relatively fast orbit to Mars. But you pretty much need fusion drive, or an equivalent array of oscillating hands, to reach Jupiter in a few months, or for practical travel to the outer planets.

Still, the Solar System as far as Jupiter should be a decent sized playground for a while.



The image comes from a NASA publication on VASIMR.